OHDSI Home | Forums | Wiki | Github

Problems with Diabetes with Renal Manfistations ICD9 mapping to SNOMED

Vocabulary Users
#1

@Christian_Reich,
I’m trying to build a concept set expression that represents diagnosis codes for Type II diabetes. In one specific case I’m having a problem trying to capture a specific ICD9 code.

The ICD9 code is 250.40: Diabetes with renal manifestations, type II or unspecified type, not stated as uncontrolled (ConceptID: 44831045). There is another type II ICD9 code: 250.42: Diabetes with renal manifestations, type II or unspecified type, uncontrolled, and what I’m trying to do is find the snomed concept that will have those ICD9 mapped over.

I found this concept: SNOMED concept Renal disorder due to type 2 diabetes mellitus (concept ID 443731) but for some reason it only maps 250.42 but NOT 250.40.

2 questions:

Do you think there’s a specific reason why 250.40 would not be considered a type of renal disorder due to type 2 diabetes?

Is it possibly a bug in the ICD9-SNOMED mapping where 250.40 is being truncated to 250.4 which is the ICD9 code for Diabetes with renal manifestations (note: it doesn’t specify the type of diabetees (1 or 2). I’m finding that the only way i can get a 250.40 code to use SNOMED concept Diabetic renal disease’ (conceptID 192279) but that is going to pull in both type 1 and type 2 diabetes ICD9s.

The reason why i propose it’s a bug, is that if I use the concept Renal disorder associated with type 1 diabetes mellitus (conceptID 200687) I actually get 350.41 and 250.43 (remember in the type2 case it shoudl be 250.40 and 250.42).

It seems that the 250.40 is just completely missed, most likely it’s confusing 250.40 and 250.4 ICD9 codes, but I just wanted to put the question to you to see if there was another more clinical reason why 250.40 is not found as a mapped concept from SNOMED: Renal disorder due to type 2 diabetes mellitus.

Any insight on this would be very very much appreciated.

#2

@Chris_Knoll:

You hit the classic example I am using when explaining the problems that can arise when mapping between incongruent hierarchies. The ICDs first determine the type of the diabetes and then define the complications, and SNOMED does it the other way around. Check it out here.

Bottom line: you will need to treat reneal disorder and diabetes separately.

#3

Ok, but don’t you think it’s strange that the snomed renal disorder for type 1 diabetes in snomed finds both the 250.41 and 250.43 but in the case of the snomed renal disorder for type 2, it doesn’t find the 250.40 but does find the 250.42 and it just so happens that it wants to map the 250.40 ICD9 up to the same concept as the 250.4 (which doesn’t distinguish between type 1 and type 2)? I strongly urge you to connfirm that there’s not some sort of truncation from 250.40 -> 250.4 happening in the mapping ICD9 -> Snomed. because this sort of error could lead to other wrong mappings.

#4

Chris: Unfortunately, you will get entangled in the vagueness and fuzziness of ICD-9 if you push it as hard as you do:

  • 250.40 says “Diabetes with renal manifestations, type II or unspecified type”. Which means it isn’t type 2. It is not clear. Which is why the mappings hooks into SNOMED further up.
  • 250.42 should actually do the same thing, but is mapped incorrectly to type 2.
  • 250.4 is the higher level concept which doesn’t go into the type, and therefore correctly maps to the same concept as 250.40 (and 250.42 if it were correct).
#5

Right, ok, so there is a problem with the mapping: 250.42 is going to the wrong place. That’s all I was looking for.

However, I’d like to make just 1 more push on this: I know that the text of the 250.40 says ‘or unspecified’ but if you read how the ICD9 is being migrated into ICD10, it’s going to E11.21 (Type 2 diabetes mellitus with diabetic nephropathy) or E11.65 which is also specifically type 2 diabetes.

Could we make a ruling here that whoever labeled that icd9 code was trying to be ‘flexible’ but we can interpret this code as a code for type two diabetes? I understand the arguments for and against, but considering how the actual ICD10 wants to interpret this code as T2, why can’t we?

Reading the details at ICD9Data.com:
http://www.icd9data.com/2015/Volume1/240-279/249-259/250/250.40.htm

Here’s the list of “Approximate Synonyms”:
Chronic kidney disease due to type 2 diabetes mellitus
Chronic kidney disease stage 1 due to type 2 diabetes mellitus
Chronic kidney disease stage 2 due to type 2 diabetes mellitus
Chronic kidney disease stage 3 due to type 2 diabetes mellitus
Chronic kidney disease stage 4 due to type 2 diabetes mellitus
Chronic kidney disease stage 5 due to type 2 diabetes mellitus
Chronic kidney disease with end stage renal disease on dialysis due to type 2 diabetes mellitus
Chronic renal impairment associated with Type 2 diabetes mellitus
Diabetes 2 with end stage renal disease on dialysis
Diabetes type 2 on dialysis
Diabetes type 2 with chronic kidney disease
Diabetes type 2 with diabetic proteinuria
Diabetes type 2 with kidney complications
Diabetes type 2 with kidney disease
Diabetes type 2 with microalbuminemia
Diabetes type 2 with mild chronic kidney disease
Diabetes type 2 with mild kidney disease
Diabetes type 2 with moderate kidney disease
Diabetes type 2 with nephrotic syndrome
Diabetes type 2 with severe kidney disease
Diabetes type 2, mild chronic kidney disease
Diabetes type 2, moderate chronic kidney disease
Diabetes type 2, severe chronic kidney disease
Diabetes, type 2 with nephropathy
Diabetic chronic renal impairment associated with Type 2 diabetes mellitus (disorder)
Diabetic stage 1 chronic renal impairment associated with Type 2 diabetes mellitus (disorder)
Diabetic stage 2 chronic renal impairment associated with Type 2 diabetes mellitus (disorder)
Diabetic stage 3 chronic renal impairment associated with Type 2 diabetes mellitus (disorder)
Diabetic stage 4 chronic renal impairment associated with Type 2 diabetes mellitus (disorder)
Diabetic stage 5 chronic renal impairment associated with Type 2 diabetes mellitus (disorder)
DM 2 W diabetic chronic kidney disease
DM 2 W diabetic chronic kidney disease, stage 1
DM 2 W diabetic chronic kidney disease, stage 2
DM 2 W diabetic chronic kidney disease, stage 3
DM 2 W diabetic chronic kidney disease, stage 4
DM 2 W diabetic chronic kidney disease, stage 5
DM 2 w diabetic CKD
DM 2 w diabetic CKD stage 1 (gfr >= 90)
DM 2 w diabetic CKD stage 1 (gfr >= 90) w HTN
DM 2 w diabetic CKD stage 2 (gfr 60-89)
DM 2 w diabetic CKD stage 2 (gfr 60-89) w HTN
DM 2 w diabetic CKD stage 3 (gfr 30-59)
DM 2 w diabetic CKD stage 3 (gfr 30-59) w HTN
DM 2 w diabetic CKD stage 4 (gfr 15-29)
DM 2 w diabetic CKD stage 4 (gfr 15-29) w HTN
DM 2 w diabetic CKD stage 5 (gfr <15)
DM 2 w diabetic CKD stage 5 (gfr <15) w HTN
DM 2 w diabetic CKD w HTN
DM 2 w diabetic CKD, ESRD on dialysis
DM 2 w diabetic CKD, ESRD on dialysis w HTN
DM 2 w diabetic microalbuminuria
DM 2 w diabetic nephropathy
DM 2 w diabetic nephrotic syndrome
DM 2 W diabetic proteinuria
DM 2 w diabetic renal manifestation
Hypertension in chronic kidney disease due to type 2 diabetes mellitus
Hypertension in chronic kidney disease stage 1 due to type 2 diabetes mellitus
Hypertension in chronic kidney disease stage 2 due to type 2 diabetes mellitus
Hypertension in chronic kidney disease stage 3 due to type 2 diabetes mellitus
Hypertension in chronic kidney disease stage 4 due to type 2 diabetes mellitus
Hypertension in chronic kidney disease stage 5 due to type 2 diabetes mellitus
Hypertension in chronic kidney disease with end stage renal disease on dialysis due to type 2 diabetes mellitus
Microalbuminuria due to type 2 diabetes mellitus
Nephrotic syndrome due to type 2 diabetes mellitus
Persistent proteinuria associated with type II diabetes mellitus
Proteinuria due to type 2 diabetes mellitus (disorder)
Renal disorder associated with type II diabetes mellitus
Stage 1 chronic renal impairment associated with Type 2 diabetes mellitus
Stage 2 chronic renal impairment associated with Type 2 diabetes mellitus
Stage 3 chronic renal impairment associated with Type 2 diabetes mellitus
Stage 4 chronic renal impairment associated with Type 2 diabetes mellitus
Stage 5 chronic renal impairment associated with Type 2 diabetes mellitus

Every single one of these is specific to Type 2.

-Chris

#6

We could.

But if we want a lex Knoll here, where do we document that exception?

#7

Hahah, ok, no i don’t want to be that way.

I should point out then that there’s some inconsistency in the vocabulary related to this case. I did some digging and I’ll give you some examples:

SNOMED concept 201826: Type 2 diabetes mellitus has mappings from the following ICD9 codes:
250.00 Diabetes mellitus without mention of complication, type II or unspecified type, not stated as uncontrolled
250.02 Diabetes mellitus without mention of complication, type II or unspecified type, uncontrolled
250.20 Diabetes with hyperosmolarity, type II or unspecified type, not stated as uncontrolled
250.22 Diabetes with hyperosmolarity, type II or unspecified type, uncontrolled

Note all the above descripts say ‘or unspecified’ so they should probably map to the generic Diabetes mellitus with hyperosmolar coma or just Diabetes Mellitus.

Another example:
SNOMED Concept 443731: “Peripheral circulatory disorder associated with type 2 diabetes mellitus” has the following ICD9s:
250.70 Diabetes with peripheral circulatory disorders, type II or unspecified type, not stated as uncontrolled
250.72 Diabetes with peripheral circulatory disorders, type II or unspecified type, uncontrolled

These say ‘or unspecified’ so should the go up to: Peripheral circulatory disorder associated with diabetes mellitus (321822)

I can go on but there are a few other cases where snomed has a group for assigning T2 + a condition, and all the ICD9s in it say ‘Type 2 or unspecified’.

So, not to keep beating this dead horse, but doesn’t it sound like 250.40 should map to Renal disorder due to type 2 diabetes mellitus just like all the other cases or should we make all the other cases behave like 250.40?

#8

All right. We’ll fix it. Next release.

#9

Thank you thank you thank you!

#10

Oh well, ICD9CM contains a lot of funny things,
Diabetes with renal manifestations, type II or unspecified type, not stated as uncontrolled, the idea is that in most cases it’s type 2, so it’s really high probability that the patient has type 2 in this case. So we may violate mapping rule and make mapping to more precise concept (like in 250.42 case). It’s interesting that we had this mapping until 2014-09-30. Seems when I reviewed this I was confused and delete this “Renal disorder due to type 2 diabetes mellitus” been afraid of violating “use more general” rule.

t